Re: [Nagios-devel] MAX_PLUGINOUTPUT_LENGTH
Posted: Fri May 26, 2006 6:39 am
Andreas Ericsson wrote:
> Goetz Rieger wrote:
>> Hi all,
>>
>> I need a longer maximum plugin output length in Nagios 2.3.1 because I
>> have a custom plugin which could possibly return quite a lot of data.
>>
>> In include/objects.h MAX_PLUGINOUTPUT_LENGTH is set to 332.
>>
>> So I have two questions:
>>
>> -Is that the correct define to alter the length of a plugin output
>> read by
>> Nagios?
>>
>
> Yes.
>
>> -Is there any drawback when setting the value rather high, say 1000?
>>
>
> Possibly, but not necessarily. The maximum posix-guaranteed atomic
> write-size is 512 bytes. If you increase this number you *might* see
> plugin output getting mixed with output from other plugins. Also, if
> you do this on a large monitoring installation you will double the
> risk of filling up the FIFO nagios uses for receiving check-commands,
> which will cause a build-up of nagios processes that will eventually
> starve the system of memory.
>
By doing, I'd tell you, if you do change this, as I've done, your
system will starve, ours has.
[]s.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
> Goetz Rieger wrote:
>> Hi all,
>>
>> I need a longer maximum plugin output length in Nagios 2.3.1 because I
>> have a custom plugin which could possibly return quite a lot of data.
>>
>> In include/objects.h MAX_PLUGINOUTPUT_LENGTH is set to 332.
>>
>> So I have two questions:
>>
>> -Is that the correct define to alter the length of a plugin output
>> read by
>> Nagios?
>>
>
> Yes.
>
>> -Is there any drawback when setting the value rather high, say 1000?
>>
>
> Possibly, but not necessarily. The maximum posix-guaranteed atomic
> write-size is 512 bytes. If you increase this number you *might* see
> plugin output getting mixed with output from other plugins. Also, if
> you do this on a large monitoring installation you will double the
> risk of filling up the FIFO nagios uses for receiving check-commands,
> which will cause a build-up of nagios processes that will eventually
> starve the system of memory.
>
By doing, I'd tell you, if you do change this, as I've done, your
system will starve, ours has.
[]s.
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]